6
$\begingroup$

I responded to this post regarding EM vs. Deep Learning, with an answer, which is to the best of my knowledge a textbook answer to the question (Full disclosure, I have a solid background in neural networks, but not so much when it comes to EM and MLE approximations).

The OP and another community member both disputed my answer. Although I still think my answer is correct, I don't want to mislead the OP and the community in case it isn't, and I was hoping that someone with a more advanced background in probability theory and learning theory could review my answer.

My deeper question is: Is it appropriate to ask for this type of review at all, or should I just delete the answer and hope someone else with the appropriate background will step in?

| |
$\endgroup$
  • 6
    $\begingroup$ This kind of reviews happen, or at least should happen, automatically via votes & comments, where each user (with some minimum number of reputation points) can participate. That is the point of sites like this. What if some user reviewed your answer and said it is correct? Would this invalidate comments by other users? Unless question gains really poor attention, usually flaws in the answers are pointed by users and wisdom of the crowds guards us from such problems. $\endgroup$ – Tim Feb 4 at 21:13
  • $\begingroup$ @Tim "What if some user reviewed your answer and said it is correct?" I understand the concern, but I would also hope that others uses might provide constructive criticism as opposed to just voicing approval or disproval. $\endgroup$ – Skander H. Feb 4 at 21:18
  • 2
    $\begingroup$ @Tim's comment effectively answers this in totality and could be turned into an answer. Otherwise put, posting anything is tacitly laying yourself open to what the rest of the community thinks, with the distinct possibility that no one wants to say anything about it either. But there isn't a special process for asking for extra views. In principle, you can do that in precisely this way, by asking a question in Meta about a question in main, but in practice you need a really special reason to do that convincingly. $\endgroup$ – Nick Cox Feb 10 at 15:28
1
$\begingroup$

Let me offer a different option: Ask a new question (on the main site) referencing the relevant question context and your answer, and asking interested respondents to explain what might be incorrect, or what might be improved. Then I would delete the original, if you aren't confident that it is correct. If you wanted, you could leave a comment on the original thread alerting readers of a relevant discussion. This provides a framework for you to get fuller and better quality information about your post, and removes a source of ambiguity and confusion from the original thread.

| |
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .